AI-designed heat pumps consume less energy
作者:Laure-Anne Pessina,洛桑联邦理工学院
Image courtesy of Ecole Polytechnique Fédérale de Lausanne
In Switzerland, between 50% and 60% of new homes are equipped with heat pumps. These systems absorb thermal energy from the surrounding environment, such as from the ground, air, or a nearby lake or river, and convert it into heat for the building.
While today's heat pumps generally work well and are environmentally friendly, they still have a lot of room for improvement. For example, by replacing traditional compression systems with micro-turbine compressors, engineers can reduce the power requirements of heat pumps by 20 to 25 percent (see illustration) and their environmental impact. This is because turbocompressors are more efficient and ten times smaller than piston units. However, it is not easy to integrate these miniature components into the design of a heat pump. Its tiny diameter (<20 mm) and fast rotational speed (>200,000 rpm) can cause complications.
At the Applied Mechanical Design Laboratory at the Ecole Polytechnique Fédérale de Lausanne on the Microcity campus, a team of researchers led by Jürg Schiffmann has developed a method to add turbocompressors to heat pumps more easily and quickly. Using a machine learning process called symbolic regression, the researchers came up with simple equations for quickly calculating the optimal size of a turbocompressor for a given heat pump. Their research just won the Best Paper Award at the Turbo Expo 2019 conference hosted by the American Society of Mechanical Engineers.
1,500 times faster
The researchers' approach greatly simplified the first step in designing a turbocharger. This step involves a rough calculation of the ideal size and speed of the required heat pump, which is important because a good preliminary estimate can significantly reduce the overall design time. Until now, engineers have been using design charts to size turbocompressors, but these charts are becoming less and less accurate as the equipment scales down. And the charts are not keeping up with the latest technology.
Image courtesy of Ecole Polytechnique Fédérale de Lausanne
That's why two PhD students at EPFL – Violette Mounier and Cyril Picard – are committed to developing an alternative. They fed the results of 500,000 simulations into a machine learning algorithm and generated equations that replicated the graph, but with several advantages: they were reliable even with small-sized turbocompressors; They are as detailed as more complex simulations; They are 1,500 times faster. The researchers' approach also allowed engineers to skip some steps in the traditional design process. It paves the way for easier implementation and wider use of micro-turbochargers in heat pumps. Compiled by Chen Jiaoyun
Advantages of micro turbo compressors
Conventional heat pumps use pistons to compress a fluid called a refrigerant and drive a vapor compression cycle. The piston needs to be well lubricated to function properly, but the oil can stick to the heat exchanger walls and impair the heat transfer process. However, a micro turbocompressor with a diameter of only a few tens of millimeters can operate without oil; They rotate at hundreds of thousands of rpm on gas bearings. The rotational motion between the components and the gas layer mean that there is almost no friction. As a result, these miniature systems can increase the heat transfer coefficient of the heat pump by 20 to 30 percent.
This micro turbocharger technology has been in development for several years and is now mature. "Several companies have already contacted us with interest in using our approach," Schiffmann said. Thanks to the researchers' work, it will be easier for companies to integrate microturbocharger technology into their heat pumps.
More information: Mounier Violette et al., Data-Driven Predesign Tools for Refrigerated Small and Medium-sized Centrifugal Compressors, Journal of Gas Turbine & Power Engineering (2018). DOI： 10.1115/1.4040845